Browse categories
Explore
Fiverr Pro
English
$
USD
I will build k6 performance tests for your backend
QA Automation Engineer
Most QA engineers add more tests. I fix the architecture.
QA Automation Engineer, Playwright specialist – 6 years across SaaS teams. One client: 45-min CI pipeline to 15 min, flake rate from 6% to un...
About this Gig
Your system works for 10 users. Production traffic doubles next month. Nobody knows what breaks first.
Most teams find out at 2am. An alert fires. Latency triples. Revenue is on fire. The post-mortem says "we should have load-tested this".
My fix: I build a k6 performance suite that simulates real production load, finds the bottleneck before launch, and runs in CI on every release.
Not a one-off script. A repeatable suite with thresholds, scenarios, and dashboards your team owns.
WHAT'S INCLUDED
- k6 project scaffolded for your repo
- Smoke, load, stress, spike, and soak scenarios
- Threshold rules that fail the build on regression
- CI integration (GitHub Actions or GitLab)
- CLI report plus optional Grafana dashboards
- Run guide for your team
WHY ME
Six years on SaaS teams. 45-min CI cut to 15 min. Built performance gates into release pipelines so a slow query never reaches prod silent. I do not run k6 once and disappear. I leave a regression net behind.
MESSAGE ME FIRST
Send your stack, target endpoints, and traffic estimate. One reply: yes I can help, or no and why. No pitch.
Testing application:
API
Development technology:
Go
•
JavaScript
Device:
Other
My Portfolio
FAQ
Why k6 instead of JMeter or Locust?
k6 is JavaScript, dev-friendly, integrates with CI cleanly, and produces metrics Grafana likes. JMeter is XML-heavy, slow to maintain, and no team I have worked with kept it alive past launch. k6 fits a modern engineering team’s actual workflow.
What traffic levels can you simulate?
Up to 10,000 virtual users on local k6 with reasonable hardware. Higher than that, we use k6 Cloud or distributed runners. I will recommend the right setup based on your target traffic on the first message.
What metrics will I get?
Response time (p50, p95, p99), error rate, throughput (requests per second), and any custom metric you care about (e.g., checkout-completed-rate). Thresholds turn these into pass/fail gates in CI.
Do you test only REST APIs?
REST is default. I also support GraphQL, WebSocket, and gRPC. Browser-based performance (real-user load) is a different gig: see my Playwright automation gig for E2E speed work, not load.
Will the tests run against production?
Smoke and small load: usually yes, off-peak. Stress and spike: only against staging or a dedicated perf environment. We agree on traffic targets and timing in the first message. I never run a stress test against live without explicit written approval.
What if you find a bottleneck?
You get a written report with the failing endpoint, the metric that failed, and likely root cause (DB, network, app code, queue). I do not fix the bottleneck in this gig. I find it. Fixing it is your engineering team or a separate engagement.
Can the tests run on every PR?
Smoke yes. Full load no, that is too expensive for every PR. Recommended setup: smoke on every PR, full load nightly, stress monthly. The CI workflow I deliver supports all three modes from day one.
What about authentication?
Token-based, OAuth, session cookies, all supported. I configure auth once in a setup script. Tests reuse the token without re-authenticating per request, which is what you want for accurate load numbers.
Can my team maintain the suite after delivery?
Yes, by design. The README explains how to add a new scenario, how to change thresholds, and how to debug a failure. Premium includes a live handover call so the team has a working session before they own it.
What if my codebase has no existing performance tests?
Most do not. The audit-style first scenario in Basic gives you a baseline number. Standard and Premium build on that baseline. You do not need a performance culture before this gig. You build one with this gig.
